http://billiards.colostate.edu/physics/Alciatore_pool_physics_article.

pdf

Pool and Billiards Physics Principles by Coriolis and Others

David G. Alciatore, PhD, PE (Dr. Dave)

Department of Mechanical Engineering
Colorado State University
1374 Campus Delivery
Fort Collins, CO 80523

David.Alciatore@colostate.edu

Abstract
The purpose of this paper is to provide illustrations and explanations of many important pool
and billiards physics principles. The goal is to provide a single and complete resource to help
physics instructors infuse billiards examples into their lectures. The main contributions of
Coriolis in his 1835 billiards physics book are presented along with other more recent
developments and experimental results. Also provided are numerous links to pool physics
references, instructional resources, and online video demonstrations. Technical derivations and
extensive experimental results are not included in the article, but they are all available and easy to
find online with the references and links provided.

key words: billiards, pool, physics, Coriolis, collision, friction, squirt, swerve, throw

I. INTRODUCTION
Pool (pocket billiards) is a great physics-teaching tool. It involves many physical principles
including conservation of momentum and energy, friction, elastic and inelastic collisions,
translational and rotational equations of motion, solid mechanics, vibrations, etc. Also, practically
all students have either played or watched pool before, so they can relate to and get excited about
pool examples, especially if the physics understanding might actually help them play better.
Furthermore, because the playing surface of a pool table is ideally flat and the balls are ideally
perfectly round and homogeneous spheres of equal mass, and ball collisions are nearly elastic and
nearly friction-free, equations written for ball trajectories can actually be solved analytically, with
only a few idealized assumptions.
In 1835, Gaspard-Gustave Coriolis wrote a comprehensive book presenting the physics of
pool and billiards. Coriolis was not only a great mathematician and physicist ... he was also an
avid billiards enthusiast. Coriolis billiards physics book has not been very widely read because it
was written in French, and an English translation has become available only recently (in 2005 by
David Nadler [1]). There has also been many technical papers and online material published over
the years expanding on pool physics knowledge [2-5]. In this article, I want to give an overview
of many of the important and useful principles that can be used as examples in physics classes. To
keep this paper of reasonable length, many of the technical derivations are provided online. I also
plan to write more papers in the future that will delve more into some of the technical details and
experimental results related to some of the principles.
The paper begins with some basic pool terminology, including important effects that come
into play when sidespin is used (i.e., when the cue ball is struck left or right of center). Then I
summarize many of the important principles discovered by Coriolis in the early 1800s. Finally, I

1
present several topics for which I have done much work over the years: the 90 and 30 rules, cue
ball squirt or deflection, and friction throw effects.
Supporting narrated video (NV) demonstrations, high-speed video (HSV) clips, and technical
proofs (TP) referenced throughout the article can be accessed and viewed online at
http://billiards.colostate.edu. The reference numbers used in the article (e.g., NV 3.1 and TP
A.6) help you locate the resources on the website; and in the online article, the references are
active links.

II. TERMINOLOGY
Figure 1 illustrates most of the important terms used to describe pool shots. The cue stick
hits the cue ball (CB) into an object ball (OB). After collision, the CB heads in the tangent line
direction and the OB heads in the impact line (AKA line of centers) direction. The key to aiming
pool shots is to be able to visualize the ghost ball (GB) target. This is the where the CB must be
when it collides with the OB to send the OB in the desired direction, which is along the line
connecting the centers of the GB and OB (see NV 3.1 and NV 3.2 for demonstrations). The cut
angle is the angle between the original CB direction (i.e. the aiming line) and the final OB
direction (i.e., the impact line).

imaginary
ghost ball (GB) tangent
target line

aiming
line
cue cue
stick ball cut
(CB) angle
object
ball line of centers
(OB) (impact line)

Figure 1 Pool terminology

Figure 1 is what happens ideally. Unfortunately, in real life, several non-ideal effects come
into play. As shown in Figure 2, when using sidespin (also known as english), where the CB is
struck to the left or right of center, the CB squirts away from the aiming line (see NV 4.13 and NV
A.17), swerves on its way to the OB (see NV 4.14 and NV 7.12), and throws the OB off the
impact line direction on its way to the target (see NV 4.15, NV 4.16, and NV A.21). If squirt,
swerve, and throw didnt exist, pool would be a much easier game to master, but the physics
wouldnt be as interesting.

2
 z y throw
squirt thrown OB angle
angle direction
x initial
CB direction
impact
line

right
bottom-right curved english (spin)
aiming
english CB path on CB
line
(swerve)

Figure 2 Non-ideal sidespin effects

Squirt and throw will be described in detail in Sections V and VI. Swerve is caused by the
fact that the cue stick is always elevated some (i.e., the back end of the cue is higher than the tip
end of the cue) to clear the rails bordering the table. Because of this, when you hit the CB off
center, the ball acquires two spin components (see Figure 3) in addition to any top or bottom spin.
One component is sidespin (about the z axis) caused by the moment created by the cues
horizontal component of impact force (Fy) about the vertical (z) axis of the CB. American pool
players refer to this component as english; interesting, the British refer to it as side. Pure
sidespin has no effect on the path of the CB until it hits a rail cushion (see NV 4.10 and NV 4.11).
The other spin component, about the horizontal aiming-line (y) axis, is called mass spin. It is
caused by the downward component of the cue sticks impact force (Fz), which creates a moment
about the y axis. HSV A.127 shows a good example of the direction and effect of mass spin.
The component of the friction force between the CB and cloth caused by the mass spin
component is perpendicular to the CBs direction of motion; hence, the CBs path will be curved.
With more cue stick elevation, the effect of the mass spin is greater causing the CBs path to
curve more (e.g., see NV 7.11, NV B.41, and NV B.42).

3
 y

A
sidespin
(wz) R x

~
~ F : resultant impulse from the cue
Fy
R: resting point on the table
top view A: impact line aim point on the table

z ~ z
Fz ~
F

masse spin
(wy) x y

R A R A
front view side view

Figure 3 English and masse spin components

III. CORIOLIS, THE MATHEMATICIAN, PHYSICIST, AND ... BILLIARDS EXPERT

Below is a concise and illustrated summary of some of the important pool physics discoveries
in Coriolis 1835 book [1]. Additional information can be found in my series of articles
describing and illustrating Coriois work [6].

4
1. The curved path followed by a CB after impact with an OB, due to top or bottom spin, is
always parabolic.

Figure 4 shows the effect of both bottom spin and topspin. If the CB had no spin, it
would travel in the tangent-line direction (per the 90 rule in Section IV). When the CB
deflects off the OB with any spin (e.g., a rolling CB would have topspin), the spin axis
remains close to its original direction immediately after impact. This is a result of
conservation of angular momentum, assuming for now friction between the balls is a small
effect (which is the case, as shown in TP A.6). The spin now creates a friction force between
the ball and cloth that has a component perpendicular to the direction of travel (see Figure
3), which causes the CBs path to curve. Because the magnitude of the sliding friction force is
nearly constant [7], the magnitude of acceleration will also be nearly constant. It turns out
that the relative velocity vector defining the direction and magnitude of the sliding motion,
and therefore the friction force vector, also does not change direction during the sliding (see
the derivation of Equation 10 in TP A.4). The relative slip speed gradually slows to zero and
remains zero thereafter (i.e., the cue ball starts rolling without slipping at a certain point and
continues to roll in a straight line, gradually slowing due to rolling resistance). Because the
friction force vector, and therefore the cue ball acceleration, are constant in both magnitude
and direction, the cue ball trajectory will be parabolic, just as with any constant acceleration
motion (e.g., projectile motion), until the sliding ceases, in which case the CB heads in a
straight line. TP A.4 contains the full derivation. HSV A.76 contains infrared super-slow
motion video of example billiard shots clearly showing parabolic traces on the cloth (caused
by heat generated by friction created by the spinning and sliding ball). Figure 5 shows a still
image from the video clearing showing a parabolic hot trace on the cloth. Also notice the
hot spots caused by the spinning CB hopping several times after being struck with a
downward stroke. In the video, you can clearly see how the sliding friction force direction
remains constant as the hot diameter develops, intensifies, and persists on the CB.

tangent
parabolic line
fast cue ball
medium paths
slow

fast
medium
slow

aiming line
impact
line

Figure 4 Parabolic CB paths

5
 Figure 5 Parabolic CB paths (Courtesy of www.bskunion.at)

2. To safely achieve maximum sidespin, the point of contact of the cue tip with the CB should be
half a ball radius off center (see Figure 6).

Obviously, the farther you hit the CB off center, the more sidespin you impart. Although,
if the tip offset exceeds the half-ball-radius (0.5R) amount, a miscue, where the cue tip slides
off the CB during impact, is very likely (see NV 2.1 and HSV 2.1). Needless to say, a miscue
is undesirable in a game situation, because the cue ball does not head in the intended
direction. Alciatore (in part IV of [6]) performed some high-speed video analysis to
experimentally determine maximum effective tip offsets possible (without miscue) with typical
pool equipment. The largest measured effective offset was about 0.55R. As shown in TP 2.1,
0.55R and 0.5R correspond to required coefficients of friction between the tip and ball of
about 0.66 and 0.58 respectively. To achieve these values, the leather tip must be properly
shaped and textured and have chalk applied properly, but all serious pool players do this, so
the 0.5R limit is appropriate for most equipment.

In Coriolis book [1], there are actually several different analyses involving tip offset. In
one analysis (a similar analysis can be found in TP A.30), he shows that for a theoretical
offset greater than 0.6R, the amount of english will be reduced because the cue tip will not
slow down enough after initial impact and it will stay in contact with the cue ball for a while;
although, he assumed the cue stick is rigid with no deflection and no cue ball squirt. The
claimed result is that the cue tip would rub on the spinning cue ball creating friction after
impact, which would reduce the amount of spin. However, high-speed video analysis

6
 (described in part IV of [6]) shows that at large offsets, the cue tip deflects away from the cue
ball and does not remain in contact. Regardless of the various analyses concerning tip offset
limitations, the miscue limit defined by the maximum possible coefficient of friction appears to
be controlling factor with typical pool equipment.

contact
point

0.5R
maximum
English

Figure 6 Contact point offset for maximum sidespin

3. With a mass shot, where the CB is struck from above off-center, the final path of the CB will
be in a direction parallel to the line drawn between the initial base point of the CB and the
aiming point on the table.

The technique is illustrated in Figure 7. The final direction of the CB path is parallel to
line RA, which connects the original CB resting point (point R) to the aiming point on the
cloth (point A). Using the letters shown in the diagram, with B indicating the CB contact
point, I refer to the Coriolis mass aiming system as the BAR method (B for ball, A
for aim, and R for resting point). This technique can be useful when trying to aim mass
shots, where you need to curve the CB around an obstacle ball (e.g., see NV 7.11, NV B.41,
and NV B.42). The detailed physics and math behind the BAR mass aiming method and
resulting ball paths can be found in TP A.19. The math, physics, and geometry is fairly
complex; but, conceptually, the result makes sense based on the fact that the initial spin axis
(which is perpendicular to the RA direction) creates a friction force component that remains
in the RA direction (per the parabolic trajectory arguments above) until the sliding ceases
and rolling begins.

7
 aiming
line

contact point curved

on the cue ball resting point cue ball
(B) of the cue ball path
on the table fast
(R ) medium
aiming point slow
on the table cloth
(A) A R
B
theoretical
final direction
side view of cue ball
(in RA direction)

top view

Figure 7 BAR mass shot aiming method

4. For a CB with natural roll, the largest deflection angle the CB can experience after impact
with an OB is 33.7, which occurs at a cut angle of 28.1 (see Section IV for illustrations and
a lot more detail on why this is useful).

IV. 90 AND 30 RULES

The most important goal in pool is sending the OB into the desired pocket (i.e., making a
shot). The second most important goal is knowing where the CB will go so you can easily make
the next shot. Figure 8 illustrates one of the most important principles of pool related to this: the
90 rule (see NV 3.4, NV 3.5, and TP 3.1). It states that when the CB strikes an OB with no
topspin or bottom spin, the two balls will always separate at 90. In other words, the CB will head
exactly in the tangent line direction and persist along this line. This is true regardless of the cut
angle (see Figure 8). Note in the remainder of this paper, the CB and OBs will be assumed to all
have equal mass. This assumption is usually very close to reality at a pool table. The CB can
sometimes be heavier (e.g., with some coin operated tables, where a mechanism under the table
can automatically sort the CB from the OBs based on larger size and/or larger weight). The CB is
more typically slightly lighter than the OBs because it experiences more collisions, some abrasion
during tip impact, and abrasion with sliding on the cloth, all of which cause wear and reduced
mass.

8
 90
(a) small cut

impact
line
the deflected
90 OB and CB paths
are 90 apart
(b) medium cut
tangent
line

90

(c) large cut

Figure 8 The 90 rule

The 90 rule is a direct result of the principles of conservation of energy and conservation of
linear momentum (see TP 3.1). All of the CBs momentum in the impact line direction (see
Figure 8b) gets transferred to the OB, assuming an elastic collision. The result is that the final
paths of the balls are perpendicular. This is what happens ideally. In real life, there is a small
amount of energy loss and retention of impact-line momentum when the CB hits the OB. This
effect is often quantified with a coefficient of restitution (COR). The COR for perfect (ideal) balls
would be 1, representing 100% conservation of energy. Typical pool balls have a COR closer to
0.93. Therefore, the 90 rule is actually closer to the 85 rule (see TP A.5).
The 90 rule, assuming elastic collisions, applies exactly only for a stun shot, where the CB is
sliding without topspin or bottom spin at impact with the OB. The one exception to the 90 rule is
when the CB hits the OB perfectly squarely, with no cut angle. In this case, the CB stops
completely, transferring all of its speed to the OB. This is called a stop shot (see NV 3.6, HSV
3.1, and HSV 3.2). Notice how in HSV 3.1, to achieve a stop shot, one must often hit the CB
below center to create bottom spin, which gradually slows, so the CB has no spin at impact with
the OB.
Figure 9 illustrates the physics involved with a stop shot. As the CB slides along the cloth,
the friction force between the ball and cloth creates a torque about the balls center that gradually
reduces (decelerates) the bottom spin (see Figure 9b and HSV 3.1). If the CB were to continue to
slide, the friction force would continue to change the CBs spin, slowly building up forward roll,
as shown in Figure 9a (see also: NV B.10). Note how the CB actually decelerates (as indicated
qualitatively by the relative lengths of the straight arrows in the figure) as the topspin builds,
because the sliding friction force is in the opposite direction as the balls motion (see the friction
arrows in the figure). Figure 9c shows what happens if the CB starts out with over-spin, which
implies more topspin than the natural roll amount. The spin rate decelerates (as indicated
qualitatively by the relative shortening of the curved arrows) until rolling develops (see HSV
B.46). In this case, the CB actually accelerates because the sliding friction force is in the same
direction as the balls motion (see the figure). When the CB is rolling, there is no longer any
sliding friction, and the CB continues to roll naturally until the ball slows to a stop (due to rolling
resistance) or hits something (e.g., another ball or a rail). If the CB is struck with the cue tip at the

9
center of percussion (COP), the ball will roll immediately. For a pool ball, the COP is at 70% of
the balls height above the table surface, for a level cue (see TP 4.2). To get over-spin (e.g., as in
Figure 9c), the CB must be struck above this height, which is risky due to likelihood for miscue.
Over-spin also occurs when a rolling CB hits an OB (e.g., see HSV B.26).

stun increasing normal

(no spin) topspin roll
table
surface

friction no sliding
friction
a) stun-to-roll conversion

bottom spin decreasing stun

bottom spin (no spin)

friction

b) bottom-spin-to-topspin conversion

decreasing normal
over-spin topspin roll

friction no sliding
friction
c) overspin-to-roll conversion

Figure 9 Conversion of vertical-plane spin to normal roll (side view)

As described above, with most pool shots, the CB will be rolling by the time it reaches the
OB. The exception is where bottom spin and/or fast speeds are being used. When the CB is
rolling, the 90 rule no longer applies. What applies instead is governed by what I call the 30 rule
(see Figure 10, NV 3.8, NV 3.9, NV 3.10, NV B.43, and NV B.44). It states that when a rolling
CB hits an OB close to a half-ball hit (see Figure 11), the CB will deflect approximately 30 away
from its initial aiming line. In TP 3.3, I show the detailed physics and math behind the rule, along
with a modern proof for the numbers in Coriolis 4th conclusion (see Section III). Wallace and
Schroeder [8] and Onoda [9] also present the supporting technical background for the 30 rule.

10
 fast
medium
slow ~30 deflected
angle ( c)

aiming
line

relaxed, but firm, peace sign used

to visualize the 30 direction

Figure 10 30 rule

half-ball
hit

Figure 11 Half-ball hit

As illustrated in Figure 10 and demonstrated in NV 3.8, you can use your hand to help
visualize the 30 CB direction. As shown in Figure 12, if you form a relaxed but firm V-shape
(peace sign or victory symbol) with your index and middle fingers, the angle between your fingers
will be very close to 30. NV 3.8, NV 3.9, and NV B.44 show how to use the Dr. Dave peace sign
in practice. If you point one of the fingers in the aiming line direction, the other finger will
indicate the direction the CB will travel after impact.
Fortunately for pool players, as shown in Figure 13, the 30 rule applies over a wide range of
ball-hit fractions (see TP 3.3). The center of the range is the half-ball hit, but the CB deflection is
very close to 30 for ball-hit fractions as small as 1/4 and as large as 3/4. The exact CB deflection
angle c as a function of cut angle is given by (see Equation 38 in TP A.4):

sin( ) cos( )
c = tan 1 (1)
sin 2 ( ) + 2

5
The derivation is fairly involved and complication; but, qualitatively, the shape in Figure 13
makes sense. For a small ball-hit fraction (i.e., large cut angle), the spin axis of the ball after
impact is close to the spin axis of the original rolling CB, and very little translational speed is lost
in the collision, so there is not much sliding between the CB and table after impact, and the CBs
direction doesnt change very much (i.e., c 0 ). For a large ball-hit fraction (i.e., small cut
angle), much of the CBs translational speed is transferred to the OB, and the spin that remains on
the CB after impact, which is almost all of the original roll topspin amount, creates sliding friction
and accelerates the ball forward (as in Figure 9c above). So in this case, the CBs original
direction also doesnt change very much (i.e., c 0 ). In the center of the range, both the post-
impact translational motion (along the tangent line) and the acceleration due to friction caused by

11
the CBs rotation (as with Figures 4 and 7 above) combine to cause a deflection angle in between
the original aiming line and the tangent line. The maximum CB deflection angle happens to occur
close to a half-ball hit (30 cut angle), as derived in TP A.4. As determined by Coriolis (see also:
TP 3.3), the actual value is approximately a 0.53-ball hit, corresponding to a 34 cut angle.

Figure 12 Using your hand to visualize the 30 rule CB paths

Figure 13 Large margin of error for 30 rule

12
 Figure 14 illustrates the ball-hit-fraction range, and corresponding cut angles, to illustrate the
wide range of shots for which the 30 rule applies. With a 1/4-ball hit (see Figure 14a), the center
of the CB is aimed outside of the object-ball edge such that the projected cue-ball-path passes
through 1/4 of the OB. With a 1/2-ball hit (see Figure 14b), the center of the CB is aimed directly
at the edge of the OB such that the projected cue-ball-path passes through 1/2 of the OB. With a
3/4-ball hit (see Figure 14c), the center of the CB is aimed inside of the object-ball edge such that
the projected cue-ball-path passes through 3/4 of the OB. These three cases cover a fairly large
range of cut angles between 14 and 49 (see TP A.23). Most pool shots are in this range.

the cue ball path is deflected by

approximately 30 in each case

medium small
large cut angle cut angle
cut angle (14.5)
(30)
(48.6)
~30 ~30 ~30

top view:
1/4 1/2 3/4

front (shooters) view:

(looking along the
aiming line)
1/4 1/2 3/4
a) 1/4-ball hit b) 1/2-ball hit c) 3/4-ball hit

Figure 14 Various ball-hit fractions

The 30 rule and peace sign technique are extremely useful in practice. This is a great
example where a little knowledge of the physics of the game can be a big help. The rule can be
used to detect possible scratches (where you pocket the CB by mistake), plan carom shots (where
you deflect one ball off another into a pocket), plan ball avoidance or break-up shots, and
strategically plan CB position as you run a rack of balls. Examples of all of these types of shots
can be found on my website (e.g., see NV 3.7, NV 3.10, NV 7.2, NV 7.3, NV 7.4, NV A.1, and
NV B.46).

13
V. SQUIRT
Figure 15 shows selected stills from a high-speed-video clip that illustrates the physics
behind squirt. The full clip, which was filmed at 2000 frames/sec, can be viewed online at HSV
A.76a. Still a is just before contact. Stills b through e represent 0.0015 sec during which
the tip is in contact with the ball. This is actually quite typical for most pool shots ... the cue tip is
in contact with the CB only for about 1-2 thousandth of a second, regardless of shot speed. In still
f the tip hasnt fully recovered from the compression yet as the CB is separating. Still g is
after separation. The line and arc appearing in each still mark the initial cue stick and CB
positions. Notice how much the cue tip deflects away (down in the diagram) from its original line
of action. Also notice how much the cue tip deforms (e.g., see still d). Those leather tips, which
havent changed much in 200 years, are quite resilient. The black arrows in still c illustrate the
effect that causes squirt. While the tip is in contact with the ball, the ball starts rotating. This
rotation (counterclockwise in the diagram) pushes the cup tip down a little during contact.
Because the end of the shaft has mass, it takes force to move the end of the shaft down as the ball
rotates. An equal and opposite reaction on the CB is what causes the CB to squirt away from the
cue aiming line.

14
 Figure 15 Close-up of cue tip impact during an off-center hit

Shepard [10] and TP A.31 provide the complete derivation relating squirt angle to tip offset
and the effective endmass of the cue, which is a function of geometry and material properties of
the end of the shaft. The endmass relates to how far the transverse elastic wave travels down the
shaft (from the tip) during the brief contact time between the tip and ball. In other words, the
endmass is a measure of how much effective mass at the end of the cue comes into play due to
transverse (sideways) motion of the tip during contact with the CB. Experiments [11, 12] have
suggested that only the last 15-20 cm (6-8 in) of typical cue shafts contribute to endmass. This
is why cue manufactures have been successful with reducing squirt by using a smaller tip and
shaft diameter, using a smaller and lighter ferrule (the white plastic component between the tip and
the shaft), and drilling out the end of the shaft, all to reduce the effective endmass [13]. It would
be interesting to do more experiments to quantify and better model how mass and stiffness
distribution affect endmass, and to measure the elastic wave speed and characterize the
dependence of transverse stiffness, but we or others have not done this yet.
In the squirt analysis, the tip is assumed to remain in contact with the ball as the ball rotates
(see HSV A.76a for visual evidence of how well a chalked leather tip "grabs" the ball); otherwise,
a miscue results, where the tip slides off the CB and the CB heads in unpredictable directions.
While the tip and ball are in contact, the velocity of the tip and ball are equal at the point of
contact. This assumption, along with linear and angular impulse-momentum principles, is used in
TP A.31 to derive the following result:
5b 2
b
1
2 R R
a = tan 1 2
(2)

1 + mb + 5 1 b
me 2 R

where, as illustrated in Figure 16, a is the squirt angle, b is the amount of tip offset from the
center (i.e., the tip contact point eccentricity), R is the ball radius (1 1/8 in = 28.6 mm), mb is the
mass of the CB (usually 6 oz = 170 gm), and me is the effective endmass of the cue (usually in the
5-15 gm range). The squirt angle is very close to a linear function of tip offset, provided miscues
and partial miscues are not occurring. Cross [14] has verified this theoretical result with a series
of experiments. The linear relation makes it easy to compensate for squirt when using sidespin
[15]. Cross [14] also investigated how squirt angle changes with tip offset under conditions of cue
tip slip, but this doesnt occur with non-miscue pool shots.

me impulses

mb
R
v

Figure 16 Squirt terminology

15
 Figure 17 shows a squirt-testing robot students and I have designed and built to perform
studies with the factors that influence squirt. The machine consists of a spring-loaded carriage on
a linear rail that consistently and accurate delivers the cue at a desired speed and CB tip-contact
point. The cue is held perfectly horizontal throughout the entire stroke (i.e., the cue is not
elevated) to eliminate CB swerve as a factor. This is an important feature because swerve can
vary with cue elevation, ball speed, shot distance, and ball and cloth conditions.

Figure 17 Cue squirt testing machine

Figure 18 shows the results of an experiment to show how end-mass affects a cues squirt. In
the experiment, a mass was added to the cue at different positions. The masses used in the
experiment weighed 0.3 grams and 1.1 grams. The two curves show the end-mass effect very
clearly. With more mass added to the end of the cue, and the closer the mass is to the tip, the more
the cue ball will squirt. Also, if mass is added beyond a certain distance from the tip, it has no
effect on the amount of squirt. For the cue tested, the distance was about 18 cm (7 in). Notice
how the curves level out to the same squirt value beyond the 18 cm (7 in), regardless of the
amount of the added mass.

16
 actual data
(1.1g mass)

actual data
(0.3g mass)

added
mass

squirt
distance

Figure 18 Squirt vs. position of added mass

One more thing to note in Figure 18 is how the squirt seemed to dip a little lower for the
larger added mass, before flattening out. When we first saw this, we wondered if adding even
more mass at this point would reduce the squirt even further. If that were the case, we might have
discovered a new way to make a low-squirt cue. Unfortunately, adding more mass at that point
did not cause additional decreases in squirt.

VI. THROW
Figure 19 illustrates all of the important terminology and physics concerning throw. Throw
occurs any time there is relative sliding motion between the CB and OB surfaces at impact. In the
figure, the relative motion is a result of CB sidespin; so the resulting throw would be called spin-
induced throw (SIT). Throw can also be caused by cut angle alone, in which case it is called cut-
induced throw (CIT). In Figure 19, the right (counterclockwise) sidespin creates a sliding friction
force that pushes the OB to the left. This force is what creates the throw angle. The spin imparted
to the OB is called transferred spin. In this case, the throwing force to the left creates clockwise
(right) transferred spin on the OB. Spin transfer is an important effect with bank shots (e.g., see
NV A.21).

17
 throw
angle
(throw )
impact
line
OB
motion

OB
center
transferred
spin
OB
throwing
force
main
impact
force

CB
motion

CB
sidespin
CB (w)

Figure 19 Throw and spin transfer terminology and physics

Throw angle throw depends on the speed of the shot v, the cut angle , the amount and
direction of CB spin w, and the friction properties of the balls. The complete derivation can be
found in TP A.14, and here is the main result:
cm (v sin( ) R z )2 + ( R x cos( ) )2
am + bm e v cos( )

thro = tan 1 min , 1 7 (v sin( ) R z ) v cos( )

(v sin( ) R z )2 + (R x cos( ) )2


(3)
where am, bm, and cm are parameters used to characterize how dynamic sliding friction varies with
the relative surface speed between the CB and OB at impact.
The 1/7 term in the min function in Equation 3 results from a requirement that relative
sliding motion between the CB and OB cannot reverse direction during impact. In the limiting
case, the relative velocity between the ball surfaces goes to zero during impact (i.e., sliding ceases
during impact). The total possible friction impulse during impact is limited by this 1/7 factor (see
TP A.14). Previous treatments of pool-ball throw in the literature (e.g., Chapter 4 in [4]), have not
included the no-slip-reverse constraint in the friction model.
The exponential model for friction used in Equation 3 is based on experimental data from
Marlow (p. 245 in [3]) and Alciatore (TP B.3). The coefficient of dynamic friction between pool
balls is far from constant ... it is much lower at higher relative speeds between the ball surfaces.

18
The exponential model of friction in Equation 3 fits experimental data quite well (see TP A.14 and
TP B.3). Possible reasons for why the friction coefficient decreases with higher relative surface
speed include: surface asperities might glance over each other (and not intermesh as easily) with a
moving surface, and boundary layer effects due to entrapped and displaced air might have a larger
effect at higher speeds.
Figure 20 shows an example plot generated from Equation 3. This particular plot shows how
the amount of throw changes with cut angle for a stun shot (i.e., sliding CB in translation only) at
various speeds (slow, medium, and fast). Useful conclusions one can observe from the figure
include:
For small cut angle shots (i.e., fuller hits), the amount of throw does not vary with shot
speed, but increases with cut angle. The reason for this is the 1/7 factor described above.
At small cut angles, sliding between the balls ceases during impact, and the coefficient of
sliding friction has no effect on the total tangential impulse experienced during the
collision.
For larger cut angle shots (i.e., thinner hits), the amount of throw is significantly larger
for slower speed shots as compared to faster speed shots. This is due to the lower
coefficient of sliding friction resulting from the faster relative speed between the sliding
ball surfaces during impact. If friction did not vary with speed (e.g., if a constant were
used for m in Equation 3), there would only be a single curve in Figure 20, and the right
portion would be a horizontal, straight line.
The amount of throw decreases a little at large cut angles, but not by much (especially for
slower speed shots). Again, this is due to the sliding friction speed dependence. At
larger cut angles, the relative speed of sliding is larger for a given ball speed.
The maximum throw occurs at close to a half-ball hit (30 cut angle).
Experiments performed by Jewett [16] and Alciatore [17, 18] have verified many of these and
other trends predicted by Equation 3. Jewetts data is shown in Figure 21. The exact throw
values in Figure 20 cannot be directly compared to those in Figure 21 because the exact physical
properties of the balls, and the precise speeds used for slow, medium, and fast, are not
available. However, the point of the figures is to show that the experimental data do exhibit the
same trends predicted by the model. Other experiments (e.g., for follow and draw shots [18]) also
verify trends predicted by the model.

19
 theoretical CIT

4 slow
throw

medium
3 fast

0
0 10 20 30 40 50 60
cut angle

Figure
20 Stun shot collision-induced throw

Figure 21 Experimental stun-shot cut-induced-throw data (Courtesy of Bob Jewett)

20
 Numerous graphs and contour plots showing how throw varies for all types of shots can be
found in TP A.28 and TP B.3. Depending on the speed, angle, and type and amount of spin on
the ball, the throw amount can vary significantly. Obviously, to play pool at a high level, ones
aim must be compensated to account for throw variations with different types of shots.

VII. CONCLUSIONS
This article has presented a collection of pool and billiards principles that make for excellent
class examples of many fundamental physics topics including friction, energy, momentum,
inelastic and elastic collisions, kinematics, and equations of motion. I have used pool examples in
many engineering and physics classes over the years, and students seem to enjoy and relate to
them. Due to size constraints, I have not included derivations or much experimental data in the
article. Instead, I have referenced online resources where the details can be found. I wanted this
article to be comprehensive summary document with good illustrations and complete lists of
online and print references for educators who want to use pool and billiards examples in their
courses. In future articles, I plan to present more details concerning derivations and experimental
results for some of the more interesting and complex topics (e.g., throw and squirt). As a final
note, if you plan to assign any pool or billiards analysis questions for homework in any of your
physics classes, there is a useful summary of typical values for all important physical parameters
in [19].

Bibliography
[1] Nadler, D., Mathematical Theory of Spin, Friction, and Collision in the Game of
Billiards, an English translation of Coriolis' 1835 book,
http:/www.coriolisbilliards.com, 2005.
[2] Alciatore, D., Pool and Billiards Physics Resources, lists of and links to many pool
physics articles, books, and websites, http:/billiards.colostate.edu/physics.
[3] Marlow, W., The Physics of Pocket Billiards, Marlow Advanced Systems Technologies
(MAST), 1995.
[4] Shepard, R., "Amateur Physics for the Amateur Pool Player," 3rd edition,
http://billiards.colostate.edu/physics/Shepard_apapp.pdf, 1997.
[5] Alciatore, D., The Illustrated Principles of Pool and Billiards, Sterling Publishing, New
York, 2004.
[6] Alciatore, D., "Coriolis was brilliant ... but he didn't have a high-speed camera - Parts I-
IV," Dr. Dave's Illustrated Principles, Billiards Digest, July-December,
http://billiards.colostate.edu/bd_articles, 2005.
[7] Witters, J. and Duymelinck, D., "Rolling and sliding resistive forces on balls moving on a
flat surface," American Journal of Physics, Vol. 54, No. 1, pp. 80-83, January, 1986.
[8] Wallace, R. and Schroeder, M., "Analysis of Billiard Ball Collisions in Two
Dimensions," American Journal of Physics, Vol. 56, No. 9, September, pp. 815-819,
1988.
[9] Onoda, G., "Comment on 'Analysis of Billiard Ball Collisions in Two Dimensions',"
American Journal of Physics, Vol. 57, No. 5, pp. 476-478, May, 1989.
[10] Shepard, R., "Everything you Always Wanted to Know About Cue Ball Squirt, but Were
Afraid to Ask," http://billiards.colostate.edu/physics/Shepard_squirt.pdf, 2001.
[11] Alciatore, D., "Squirt - Part VII: cue test machine results," Dr. Dave's Illustrated
Principles, Billiards Digest, Vol. 30, No. 3, pp. 38-39, February, 2008.
[12] Alciatore, D., "Return of the squirt robot," Dr. Dave's Illustrated Principles, Billiards
Digest, Vol. 30, No. 9, pp. 36-37, August, 2008.

21
[13] Alciatore, D., "Squirt - Part V: low-squirt cues,", Dr. Dave's Illustrated Principles
Billiards Digest, Vol. 30, No. 1, pp. pp. 38-39, December, 2007.
[14] Cross, R., "Cue and ball deflection (or squirt) in billiards," American Journal of
Physics, Vol. 76, No. 3, pp. 205-212, 2008.
[15] Alciatore, D., "Squirt - Part IV: BHE, FHE, and pivot-length calibration," Dr. Dave's
Illustrated Principles, Billiards Digest, Vol. 29, No. 12, pp. 40-41, November, 2007.
[16] Jewett, B., Seeking truth of beliefs, Tech Talk, Billiards Digest, pp. 30-32, June,
1995.
[17] Alciatore, D., "Throw - Part II: results," Dr. Dave's Illustrated Principles, Billiards
Digest, Vol. 29, No. 10, September, 2006.
[18] Alciatore, D., "Throw - Part III: follow and draw effects," Dr. Dave's Illustrated
Principles, Billiards Digest, Vol. 29, No. 11, October, 2006.
[19] Alciatore, D., Approximate values for all pool/billiards-related physical properties,
http://billiards.colostate.edu/threads/physics.html#properties, 2007.

Note:
All of the Billiards Digest articles referenced above are also available at:
http://billiards.colostate.edu

22